A synthetic route to the saxitoxin skeleton: Synthesis of decarbamoyl α-saxitoxinol, an analogue of saxitoxin produced by the cyanobacterium lyngbya wollei

Article abstract of DOI:10.1002/anie.201102494

Pick your poison: Salient features of a concise synthetic route for the saxitoxin skeleton are a bromonium-ion-initiated cascade cyclization of readily prepared 1 to give tricyclic intermediate 2 and the transformation of the gem-dibromomethylene group in

Full text of DOI:10.1002/anie.201102494

Communications
DOI: 10.1002/anie.201102494
Natural Products
A Synthetic Route to the Saxitoxin Skeleton: Synthesis of Decarbamoyl
a-Saxitoxinol, an Analogue of Saxitoxin Produced by the
Cyanobacterium Lyngbya wollei**
Yusuke Sawayama and Toshio Nishikawa*
Dedicated to the researchers affected by the devastating 2011 earthquake of Japan
Saxitoxin (1), which was first isolated as a paralytic shellfish
poison,[1] is a potent and specific blocker of voltage-gated
sodium channels,[2] similar to tetrodotoxin, the puffer fish
toxin. This unique biological activity prompted us to use
saxitoxin and tetrodotoxin as biochemical tools for the study
of voltage-gated sodium channels and other ion channels in
the field of neurophysiology.[3] Recently, molecular genetic
analyses have revealed that ten subtypes of voltage-gated
sodium channels are expressed in mammals, and that these
subtypes appear to have specific functions in different
organs.[4] However, the specific functions of these channels
have yet to be clarified. Our group is interested in developing
a subtype-selective blocker of voltage-gated sodium channels
based on natural compounds, such as tetrodotoxin and
saxitoxin, and have previously established synthetic methods
for tetrodotoxin and its analogues.[5] To expand the range of
available candidate subtype-selective blockers against volt-
age-gated sodium channels, we are attempting to develop a
novel synthetic route for saxitoxin and its analogues, including
neo-saxitoxin, gonyautoxin 3, and zetekitoxin AB, which was
isolated from a Panamanian golden frog (Figure 1).[6]
Figure 1. Structures of saxitoxin (1) and its naturally occurring
analogues.
Since its structural elucidation in 1975, saxitoxin has been
an attractive target molecule for total synthesis owing to its
potent biological activity, as well as its unique densely
functionalized structure, which features a bicyclic guanidi-
nium condensed with a pyrrolidine and a hydrated form of the
C12 carbonyl group. The total synthesis of saxitoxin was first
reported by Kishi and co-workers in 1977[7] and subsequently
by Jacobi et al. in 1984.[8] During the last five years, Du Bois
and co-workers,[9] and Nagasawa and co-workers[10] have
independently studied the synthesis; these studies culminated
in the total syntheses of saxitoxin and its analogues, including
gonyautoxin 3. Herein, we describe a novel synthetic route for
the saxitoxin skeleton; this route enabled us to synthesize
decarbamoyl a-saxitoxinol (2),[11] which is a nontoxic, natu-
rally occurring analogue of saxitoxin produced by the
cyanobacterium Lyngyba wollei.
To synthesize saxitoxin (1) and its analogues, we devel-
oped a novel bromocyclization strategy for the syntheses of
cyclic guanidines from simple acyclic substrates; five- and six-
membered cyclic guanidines were formed from propargyl and
homopropargyl guanidine, respectively.[12] Scheme 1 shows an
example of the bromocyclization; this reaction yielded the
six-membered cyclic guanidine 4 possessing a spiro-aminal
structure, which is fully functionalized for the saxitoxin
skeleton. However, subsequent reduction of the azide group
of 4 in preparation for the installation of the second guanidine
functional group proved to be difficult, probably because of
the severe steric hindrance arising from the geminal dibromo
substituents. In the course of these synthetic experiments, we
incidentally found a novel reaction involving the transforma-
tion of the gem-dibromomethylene group of 4 into a carbonyl
functionality, as shown in Scheme 1. In this reaction, dibromo
spiro-aminal compound 4 was treated with acetic anhydride
and triethylamine in dichloromethane at room temperature to
afford enol acetate 5 as an unstable product,[13] and alcohol
6[14] was isolated after a reduction with NaBH4 and a
[*] Dr. Y. Sawayama, Prof. T. Nishikawa
Graduate School of Bioagricultural Sciences, Nagoya University
Chikusa, Nagoya 464-8601 (Japan)
E-mail: nisikawa@agr.nagoya-u.ac.jp
[**] We are grateful to Prof. Yasukatsu Oshima (Kitasato University,
School of Marine Biosciences) for providing NMR spectra of
compound 2 (Lyngbya wollei toxin 4). This work was financially
supported by a Grant-in-Aid for Scientific Research and a G-COE
grant from the Japan Society for the Promotion of Science (JSPS),
the Naito Foundation, and the Nagase Foundation. Y.S. thanks the
SUNBOR Scholarship and Nagoya University scholarship for
outstanding graduate students.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201102494.
7176
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7176 –7178

Scheme 1. Bromocyclization for the synthesis of the cyclic guanidine 4
and the novel transformation of gem-dibromomethylene into a carbon-
yl functionality. Reagents and reaction conditions: a) TFA, CH2Cl2, RT,
2 h; b) PyHBr3 (5 equiv), K2CO3 (10 equiv), CH2Cl2, H2O, RT 1 h, 71%
(for 2 steps); c) Ac2O, Et3N, CH2Cl2, RT, 30 min; d) NaBH4 (2 equiv),
MeOH, RT, 10 min; then K2CO3 (1.3 equiv), RT, 1 h, 64% (for 2 steps).
Boc=tert-butyloxycarbonyl, Cbz=benzyloxycarbonyl, Py=pyridine,
TFA=trifluoroacetic acid.
Scheme 3. Synthesis of 10 and its bromocyclization to give tricyclic
compound 11. Reagents and reaction conditions: a) NaN3 (1.5 equiv),
DMF, RT, 4.5 h; b) TBAF (1.5 equiv), THF, RT, 30 min; c) MsCl
(1.05 equiv), Et3N (3 equiv), CH2Cl2, 08C to RT, 40 min; d) KCN
(1.05 equiv), EtOH, RT, 12 h; e) TFA, CH2Cl2, RT, 2 h; f) PyHBr3
(3 equiv), K2CO3 (10 equiv), CH2Cl2, H2O, RT, 1 h, 24% for 6 steps.
DMF=N,N-dimethylformamide, TBAF=tetrabutylammonium fluoride,
THF=tetrahydrofuran, Ms=methanesulfonyl.
subsequent hydrolysis of the acetate with K2CO3 in methanol.
However, further efforts to synthesize the saxitoxin skeleton
from 6 were unsuccessful.
Based on the above results, we devised an alternative
synthetic route for the saxitoxin skeleton, as shown in
Scheme 2. We envisaged that the tricyclic skeleton A of
saxitoxin would be formed by a guanidine cyclization of B via
an iminium ion intermediate under acidic conditions.[15]
Precursor B could then be prepared from the tricyclic
intermediate C by the transformation of the gem-dibromo-
methylene group under the reaction conditions mentioned
above followed by the installation of the guanidine function-
ality. We planned to synthesize C by a cascade cyclization of
E, which was initiated by a bromonium ion, as the expected
oxaazabicyclo[3.2.1] product D would undergo an intramo-
lecular N alkylation to give pyrrolidine, which corresponds to
the C ring of saxitoxin. Precursor E could then be prepared
from guanidino aziridine 7, an intermediate for the synthesis
of 3, as described in our previous report.[12]
The precursor for the bromocyclization reaction (10),
which has a similar structure to E, was synthesized from 7 in
five steps, as shown in Scheme 3. Compound 7 underwent ring
opening with NaN3 in N,N-dimethylformamide at room
temperature, and the siloxy group of product 8 was trans-
formed to the corresponding mesylate 9 in two steps by
desilylation and mesylation under conventional reaction
conditions. Sequential deprotection of the acetate with
KCN in EtOH[16] and the Boc group with trifluoroacetic
acid yielded 10, which served as the precursor for the
bromocyclization. As anticipated, upon treatment of 10 with
pyridinium tribromide (PyHBr3) in a biphasic medium of
CH2Cl2 and aqueous K2CO3 at room temperature, the key
bromocyclization reaction and subsequent intramolecular
N alkylation took place to afford 11 as a single product.
Since intermediates 9 and 10 were unstable, the optimal result
was obtained when these intermediates were not purified, and
this approach allowed for the preparation of 11 in 24%
overall yield in 6 steps from 7 (average yield of 79% in each
step).
After obtaining the key intermediate 11, we next exam-
ined the transformation of the gem-dibromomethylene group
of 11 into a carbonyl functionality under the reaction
conditions used in Scheme 1. Using this approach, compound
11 was successfully transformed into alcohol 12[17] in two
steps: 1) Ac2O, Et3N in CH2Cl2 and 2) reduction of the
resulting enol acetate with NaBH4 in MeOH (Scheme 4). The
azide group of 12 was then reduced under Staudinger
conditions (PMe3 in CH2Cl2 and subsequent hydrolysis with
aqueous HCl and MeOH) followed by a conventional
Scheme 2. Synthetic plan for the saxitoxin skeleton. TBS=tert-butyl-
dimethylsilyl.
Angew. Chem. Int. Ed. 2011, 50, 7176 –7178
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7177

Communications
97, 1238 – 1239; b) J. Bordner, W. E. Thiessen, H. A. Bates, H.
Rapoport, J. Am. Chem. Soc. 1975, 97, 6008 – 6012.
[2] a) V. L. Salgado, J. Z. Yeh, T. Narahashi, Ann. N. Y. Acad. Sci.
1986, 479, 84 – 95; b) L. E. Llewellyn, Nat. Prod. Rep. 2006, 23,
200 – 222.
[3] a) F. Hucho, Angew. Chem. 1995, 107, 23 – 36; Angew. Chem. Int.
Ed. Engl. 1995, 34, 39 – 50; b) Tetrodotoxin, Saxitoxin, and the
Molecular Biology of the Sodium Channel (Eds.: C. Y. Kao, S.
Lovinson), Ann. N. Y. Acad. Sci. 1986, 479.
[4] A. L. Goldin, R. L. Barchi, J. H. Caldwell, F. Hofmann, J. R.
Howe, J. C. Hunter, R. G. Kallen, G. Mandel, M. H. Meisler,
Y. B. Netter, M. Noda, M. M. Tamkun, S. G. Waxman, J. N.
Wood, W. A. Catterall, Neuron 2000, 28, 365 – 368.
[5] a) D. Urabe, T. Nishikawa, M. Isobe, Chem. Asian J. 2006, 1,
125 – 135; b) T. Nishikawa, D. Urabe, K. Yoshida, T. Iwabuchi,
M. Asai, M. Isobe, Chem. Eur. J. 2004, 10, 452 – 462, and
references therein.
[6] a) Y. Shimizu, Ann. N. Y. Acad. Sci. 1986, 479, 24 – 31; b) E. J.
Schantz, Ann. N. Y. Acad. Sci. 1986, 479, 15 – 23; c) M. Yotsu-
Yamashita, Y. H. Kim, S. C. Dadley, Jr., G. Choudhary, A.
Pfahnl, Y. Oshima, J. W. Daly, Proc. Natl. Acad. Sci. USA 2004,
101, 4346 – 4351.
[7] a) H. Tanino, T. Nakata, T. Kaneko, Y. Kishi, J. Am. Chem. Soc.
1977, 99, 2818 – 2819; b) Y. Kishi, Heterocycles 1980, 14, 1477 –
1495; c) C. Y. Hong, Y. Kishi, J. Am. Chem. Soc. 1992, 114, 7001 –
7006.
[8] a) P. A. Jacobi, M. J. Martinelli, S. Polanc, J. Am. Chem. Soc.
1984, 106, 5594 – 5598; b) M. J. Martinelli, A. D. Brownstein,
P. A. Jacobi, S. Polanc, Croat. Chem. Acta 1986, 59, 267 – 295;
c) P. A. Jacobi in Strategies and Tactics in Organic Synthesis,
Vol. 2 (Eds.: T. Lindberg), Academic Press, New York, 1989,
pp. 191 – 219.
[9] a) J. J. Fleming, J. Du Bois, J. Am. Chem. Soc. 2006, 128, 3926 –
3927; b) J. J. Fleming, M. D. McReynolds, J. Du Bois, J. Am.
Chem. Soc. 2007, 129, 9964 – 9975; c) J. V. Mulcahy, J. Du Bois, J.
Am. Chem. Soc. 2008, 130, 12630 – 12631; d) B. M. Andresen, J.
Du Bois, J. Am. Chem. Soc. 2009, 131, 12524 – 12525.
[10] a) O. Iwamoto, H. Koshino, D. Hashizume, K. Nagasawa,
Angew. Chem. 2007, 119, 8779 – 8782; Angew. Chem. Int. Ed.
2007, 46, 8625 – 8628; b) O. Iwamoto, R. Shinohara, K. Naga-
sawa, Chem. Asian J. 2009, 4, 277 – 285; c) O. Iwamoto, K.
Nagasawa, Org. Lett. 2010, 12, 2150 – 2153.
[11] H. Onodera, M. Satake, Y. Oshima, T. Yasumoto, W. W.
Carmichael, Nat. Toxins 1997, 5, 146 – 151.
[12] Y. Sawayama, T. Nishikawa, Synlett 2011, 651 – 654.
[13] The generality and mechanistic studies of this reaction are
currently under investigation in this laboratory.
[14] The configuration of the C12-position was determined by
NOESY correlation between protons of the C5- and C12-
positions. See the Supporting Information.
[15] Du Bois reported the synthesis of a similar compound (ketone at
C12-position) to B, however, cyclization of the five-membered
guanidine was unsuccessful. We speculated that the hydrate
functionality (ketone equivalent) at the C12-position exhibited
resistance to the formation of an iminium ion intermediate
owing to the electron-withdrawing nature of the hydrate (see
Ref. [8b]). We anticipated that B could be a substrate for this
cyclization reaction because of the weak electron-withdrawing
nature of the hydroxy group at the C12-position.
Scheme 4. Synthesis of decarbamoyl a-saxitoxinol (2). Reagents and
conditions: a) Ac2O, Et3N, CH2Cl2, RT, 15 min; b) NaBH4 (30 equiv),
MeOH, RT, 30 min, 32% (for 2 steps); c) Me3P (3 equiv), CH2Cl2, RT,
30 min; then MeOH/12m HCl (5:1), RT, 30 min; d) CbzN=C(SMe)-
NHCbz (2 equiv), HgCl2 (2 equiv), Et3N (10 equiv), DMF, 608C, 51%
(for 2 steps); e) H2 (1 atm), 10% Pd/C, MeOH, EtOAc, RT, 27 h;
f) B(OCOCF3)3, TFA, RT, 24 h, 73% (for 2 steps).
guanidinylation to give 13. Deprotection of the benzyloxy-
carbonyl groups under hydrogenolytic conditions yielded 14,
a precursor for the next cyclization reaction. Upon treatment
with B(OCOCF3)3 in TFA at room temperature,[18] 14 under-
went ring opening of the hemiaminal and subsequent
cyclization of the second guanidine to furnish decarbamoyl
a-saxitoxinol 2 (L. wollei toxin 4), a metabolite of the
cyanobacterium L. wollei, in 73% yield. The spectroscopic
data of compound 2 synthesized in this study were in good
agreement with those of the natural product.[11]
In summary, we have developed a concise synthetic route
for the saxitoxin skeleton, featuring two key reactions:
cascade cyclization to give tricyclic intermediate 11, and
transformation of the gem-dibromomethylene group into enol
acetate. Because the hemiaminal of 11 is stable owing to the
electron-withdrawing nature of the gem-dibromo substitu-
ents, the intermediate should be suitable for the incorporation
of additional functionality, which will provide diverse ana-
logues of saxitoxin for the development of subtype-selective
blockers of voltage-gated sodium channels. Further studies
toward the synthesis of other saxitoxin analogues, such as
gonyautoxin 3 and zetekitoxin AB, employing this method are
currently under investigation in our laboratory.
Received: April 11, 2011
Published online: June 17, 2011
[16] K. Mori, M. Tominaga, T. Takigawa, M. Matsui, Synthesis 1973,
790 – 791.
[17] The configuration of the C12-position was determined by
analysis of NOESY spectra of 13. See the Supporting informa-
tion.
Keywords: cyclization · guanidines · natural products · saxitoxin ·
total synthesis
.
[18] The conditions were identical to those employed in the total
syntheses of saxitoxin by Du Bois and co-workers, and Nagasawa
and co-workers (see Ref. [9a–c and 10b].
[1] a) E. J. Schantz, V. E. Ghazarossian, H. K. Schnoes, F. M. Strong,
J. P. Springer, J. O. Pezzanite, J. Clardy, J. Am. Chem. Soc. 1975,
7178
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7176 –7178